Psychometric function with one separate data point. More...

#include <psychometric.h>

Inheritance diagram for OutlierModel:

Public Member Functions
	OutlierModel (int nAFC, PsiCore core, PsiSigmoid sigmoid, unsigned int exclude)
void	setexclude (unsigned int exclude)
	set up a psychometric function model that treats one block separately
double	negllikeli (const std::vector< double > &prm, const PsiData *data) const
	negative log likelihood
double	neglpost (const std::vector< double > &prm, const PsiData *data) const
	negative log likelihood
double	deviance (const std::vector< double > &prm, const PsiData *data) const
	deviance
unsigned int	getNparams (void) const
	get the number of free parameters of the psychometric function
double	randPrior (unsigned int index) const
	sample form a prior

Detailed Description

Psychometric function with one separate data point.

Wichmann & Hill (2001) suggest to detect whether or not a data point x0 is an outlier by fitting a new model to all data points except for the data point of interest and to use a separate parameter to fit x0. If p is the number of correct trials at position x0, that is

Psi(x0) = p Psi(x) = guessingrate + (1-guessingrate-lapserate) * Sigmoid ( x | theta ), if x != x0

This model is then compared to the standard psychometric function model.

Due to the discontinuity at x0, derivatives in this model are treated with respect to the fitted function.

Constructor & Destructor Documentation

OutlierModel::OutlierModel	(	int	nAFC,
		PsiCore *	core,
		PsiSigmoid *	sigmoid,
		unsigned int	exclude
	)			`[inline]`

Parameters:

	nAFC	number of alternatives in the task (1 indicating yes/no)
	core	internal part of the nonlinear function
	sigmoid	"external" saturating part of the nonlinear function
	exclude	index of the data block to be excluded

Member Function Documentation

double OutlierModel::deviance	(	const std::vector< double > &	prm,
		const PsiData *	data
	)			const `[virtual]`

deviance

Parameters:

	prm	parameters of the psychometric function model
	data	data for which the deviance should be evaluated

Reimplemented from PsiPsychometric.

double OutlierModel::negllikeli	(	const std::vector< double > &	prm,
		const PsiData *	data
	)			const `[virtual]`

negative log likelihood

Parameters:

	prm	parameters of the psychometric function model
	data	data for which the likelihood should be evaluated

Reimplemented from PsiPsychometric.

double OutlierModel::neglpost	(	const std::vector< double > &	prm,
		const PsiData *	data
	)			const `[virtual]`

negative log likelihood

Parameters:

	prm	parameters of the psychometric function model
	data	data for which the likelihood should be evaluated

Reimplemented from PsiPsychometric.

void OutlierModel::setexclude ( unsigned int exclude ) [inline]

set up a psychometric function model that treats one block separately

change the excluded block

The documentation for this class was generated from the following files:

src/psychometric.h
src/psychometric.cc

OutlierModel Class Reference

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation